845Development 125, 845-856 (1998)Printed in Great Britain © The Company of Biologists Limited 1998DEV4958

Sequential roles for Otx2 in visceral endoderm and neuroectoderm for

forebrain and midbrain induction and specification

Muriel Rhinn 1, Andrée Dierich 1, William Shawlot 2, Richard R. Behringer 2, Marianne Le Meur 1

and Siew-Lan Ang 1,*1Institut de Génétique et de Biologie Moléculaire et Cellulaire, CNRS/INSERM/Université Louis Pasteur, 67404 Illkirch cedex, CUde Strasbourg, France2Department of Molecular Genetics, The University of Texas M. D. Anderson Cancer Center, Houston, Texas 77030, USA*Author for correspondence (e-mail: siew-lan@igbmc.u-strasbg.fr)

Accepted 12 December 1998: published on WWW 4 February 1998

The homeobox gene Otx2 is a mouse cognate of theDrosophila orthodenticle gene, which is required fordevelopment of the brain, rostral to rhombomere three. Wehave investigated the mechanisms involved in this neuralfunction and specifically the requirement for Otx2 in thevisceral endoderm and the neuroectoderm using chimericanalysis in mice and explant recombination assay. Analysesof chimeric embryos composed of more than 90% of Otx2−/− ES cells identified an essential function for Otx2 in thevisceral endoderm for induction of the forebrain andmidbrain. The chimeric studies also demonstrated that ananterior neural plate can form without expressingOtx2.

However, in the absence of Otx2, expression of importantregulatory genes, such as Hesx1/Rpx, Six3, Pax2, Wnt1andEn, fail to be initiated or maintained in the neural plate.Using explant-recombination assay, we could furtherdemonstrate that Otx2 is required in the neuroectodem forexpression of En. Altogether, these results demonstrate thatOtx2 is first required in the visceral endoderm for theinduction, and subsequently in the neuroectoderm for thespecification of forebrain and midbrain territories.

Key words: Otx2, forebrain, midbrain, patterning, chimera, ES cell,explant, orthodenticle, Drosophila, mouse

SUMMARY

er

d

rall.,),iore

es

drlyrm in

5;te

eof

INTRODUCTION

The molecular mechanisms that control regional specificatof the vertebrate brain are largely unknown, although magenes homologous to Drosophilapattern-formation genes areexpressed in the embryonic brain and are candidates playing a role in its regionalization. Genetic analysis of hedevelopment in Drosophila has identified three gap genesorthodenticle, empty spiraclesand buttonhead, that areimportant for the development of head segments (reviewedCohen and Jurgens, 1991; Finkelstein and Perrimon, 199Otx1, Otx2 (Simeone et al., 1993), and Emx1 and Emx2(Simeone et al., 1992) are cognate genes of otd and emsgenes,respectively, in mice. These genes show overlappiexpression patterns in the forebrain and midbrain. Other gesuch as Six3(Oliver et al., 1995), a homologue of sine oculis,Pax2(Rowitch and McMahon, 1995), a homologue of paired,and En1 and En2 (Davis and Joyner, 1988), homologues oengrailed, also show restricted expression domains in tforebrain, midbrain and/or anterior hindbrain of early mouembryos. These genes encode paired domain or homeodomproteins that are likely to be transcription factors. Therestricted expression patterns along the anteroposterior (Aaxis of the neural tube suggest that they may control stepwise regionalization of the brain along this axis.

Otx2 is the earliest homeobox-containing gene to b

ionny

forad,

in1).

ngnes

fhese

ainir-P)

the

e

expressed in the neuroectoderm. Otx2 is a bicoid-classhomeobox gene, which has also been isolated from othspecies, including sea urchin (Gan et al., 1995), Xenopus(Pannese et al., 1995), chick (Bally-Cuif et al., 1995) anzebrafish (Li et al., 1994, Mercier et al., 1995). In mouse, Otx2is already expressed in the entire ectoderm and visceendoderm before gastrulation at E5.5-6.0 (Simeone et a1993; Ang et al., 1994). As gastrulation proceeds (E6.0-E7.5this expression becomes progressively restricted to the anterend of the embryo in all three germ layers. In vitro tissurecombination experiments in mouse and Xenopus havedemonstrated that interactions with axial mesendoderm tissuare involved in the progressive restriction of Otx2 expressionto the anterior end of the embryo (Ang et al., 1994; Blitz anCho, 1995). In mice, this restriction starts as soon as the eagastrulation stage (E6.5), suggesting that visceral endodesurrounding the epiblast at this stage could also be involvedthis restriction.

To study the function of Otx2during mouse embryogenesis,we and others have generated mutations in Otx2 byhomologous recombination in ES cells (Acampora et al., 199Matsuo et al., 1995; Ang et al., 1996). Homozygous mutanembryos show early gastrulation defects, in particular in thdevelopment of axial midline cells, and they fail to developforebrain, midbrain and anterior hindbrain. However, thesstudies did not pinpoint the mechanism leading to the loss

846

og3. kb

tor

outas

al.,m

ify

e

tte,(te of

on

31

ereenes.

ypeS

and

ws.d

d in

o-

ede

asthe

M. Rhinn and others

rostral brain tissue. In vivo transplantation and in vitro tissrecombination experiments, pioneered by Spemann aMangold (reviewed in Lemaire and Kodjabachian, 1996), hademonstrated that an interaction between ectoderm and amesendoderm plays an important role in the formation aregionalization of the neural tube. Since Otx2 is expressed inboth the ectoderm and the axial mesendoderm at the anteend of the embryo at the time of neural plate formation (Aet al., 1994), it could be required in either of the two tissufor rostral brain development. Otx2 could be required inmesendodermal cells for the production of signals inducing anterior neural plate formation. Alternatively, loss of rostrbrain tissue could be due to an autonomous function of Otx2in the ectoderm and/or neurectoderm.

A third possibility is that Otx2 is required in the visceralendoderm for development of the rostral brain. Viscerendoderm derives from the primitive endoderm and, thus, a different origin from embryonic endoderm that gives rise the gut and is derived from the epiblast only later, at E6During gastrulation, the visceral endoderm joins with thextraembryonic mesoderm to form the visceral yolk s(VYS). The critical role of the VYS in the maternofetaexchange of nutrients has been well documented (Jollie, 19Visceral endoderm cells have also been shown in getargeting experiments to play an important role in the growof the epiblast (Chen et al., 1994; Spyropoulus et al., 19Duncan et al., 1997). In addition, recent embryological achimeric studies in mice have provided evidence for a rolevisceral endoderm in patterning the rostral brain durigastrulation (Thomas and Beddington, 1996; Varlet et a1997).

In this work, we aimed at better understanding the functiof Otx2 in anterior neural tube development, revealed by tloss of rostral brain tissue in Otx2−/− embryos. We first showthat the anterior neural plate is missing in these embryos atearliest stage identifiable. These results strongly suggest the loss of rostral brain in Otx2−/− embryos is due to a lack ofinduction of the anterior neural plate, which normally requirectoderm-mesendoderm interactions. To determine in whgerm layers Otx2 is required for the induction andregionalization of the rostral brain tissue, we performechimera studies with homozygous Otx2−/− ES cells and in vitrotissue recombination experiments between Otx2−/− and wild-type embryos. Chimeric embryos composed of more than 9Otx2−/− cells show a rescue of the anterior neural plate, duethe presence of wild-type cells in the visceral endoderm. Trescued rostral neural tissue, which consists entirely of Otx2−/−

cells, however fails to express some forebrain and midbrgenes. In addition, Otx2−/− ectoderm did not express En1 anEn2 (En) proteins after recombination with wild-typmesendoderm in vitro. Altogether, these results demonstrthat Otx2 is required in the visceral endoderm for anterioneural plate induction, and subsequently in the neuroectodfor regional specification of the forebrain and midbrain.

MATERIALS AND METHODS

Targeting vectorA 1 kb Otx2 cDNA (Ang et al., 1994) was used to isolate from 129Sv/J genomic library two overlapping genomic clones, g3 and

uendvexialnd

riornges

theal

alhasto.5.e

acl90).neth

94;nd ofngl.,

onhe

thethat

esich

d

0% tohis

aindeater

erm

a g1

(Fig. 1A), encoding the entire coding region of the Otx2 gene. TheNotI-NotI insert of genomic clone, g3 was subcloned intpBluescriptIIKS (+) (Stratagene), to generate a plasmid called BSThe g3-hygro targeting vector was constructed by replacing the 4BglII fragment of plasmid BSg3 with a 2 kb ClaI-XhoI fragmentcontaining a PGKhygro cassette, using blunt-end ligation. This veccontains 8.5 kb of 5′ and 1.5 kb of 3′ homology (Fig. 1A).

Generation of Otx2 mutant and wild-type ROSA26 ES celllinesThe culture, electroporation and selection of ES cells were carried as described (Dierich and Dollé, 1997). The g3-hygro construct welectroporated after linearisation by SalI into the ES cell line TE123heterozygous for the Otx2neomutation, whereby the homeodomain ofOtx2 was deleted and replaced by the PGKneo cassette (Ang et1996). Electroporated cells were placed in a selective mediucontaining 150 µg/ml of hygromycin B. ES cell colonies that wereresistant to hygromycin were analysed by Southern blotting to identhomologous recombination events at either the wild-type or the neoallele of the Otx2 locus. Homologous recombination between thtargeting vector g3-hygro and the wild-type or Otx2neo allele replacesthe first and second exons of Otx2, including the ATG translation startsite and most of the homeodomain or the PGKneo casserespectively, with the PGKhygro cassette from the pHA58 vector Riele et al., 1990), in a transcriptional orientation opposite to thatthe Otx2 gene, resulting in a new mutant allele of the Otx2gene calledOtx2hygro (Fig. 1A). Genomic DNA from ES cells was digested witheither EcoRI or NcoI and probed with a 4.0 kb XbaI-XbaI 3′ flankingfragment and with a 1.0 kb XhoI-XhoI 5′ flanking fragment,respectively (Fig. 1A). Of 144 colonies analyzed, 4 Otx2neo/hygro

(Otx2−/−) ES cell lines called TX3, TX62, TX81 and TX112 and 5Otx2hygro/+ (Otx2+/−) ES cell lines called TX2, TX93, TX106, TX107and TX111 were obtained, resulting from homologous recombinatievents occurring into the wild-type or Otx2neoallele, respectively (Fig.1B and data not shown). The wild-type ROSA26/+ ES cell line, ESwas isolated from blastocysts obtained from Lim1−/+, ROSA26/+intercrosses, as described (Robertson, 1987).

Generation and analysis of chimerasThe morulae-stage (E2.5) embryos used to generate chimeras wobtained from crosses of males homozygous for the ROSA 26 gtrap insertion (Friedrich and Soriano 1991) with CD1 femaleEmbryos were injected with approximately ten Otx2−/− (TX3, TX62,TX112), Otx2+/− (TX106 and TX111) or wild-type ES cells (R1 andH1) and reimplanted into pseudogestant females. The reciprocal tof chimeras were generated by injecting wild-type ROSA26/+ Ecells (ES31) into morulae isolated from Otx2+/− intercrosses.Chimeric embryos were harvested between E8.5 and E9.5, processed for in situ hybridization or β-galactosidase staining asdescribed in Beddington et al. (1989). After β-galactosidase staining,some embryos were processed for histological analysis as folloStained embryos were postfixed overnight in Bouin’s fixative, rinsein 70% ethanol, dehydrated and embedded in paraffin as describeKaufman (1992). 7 µm thick sections were counterstained with 0.01%safranin-O that stained ES-derived cells in pink whereas embryderived cells were blue. In experiments using embryos from Otx2+/−

intercrosses, the genotype of the host morula was determinretrospectively from VYS endoderm, that was isolated from thmesoderm layer following digestion in pancreatin and trypsin described (Hogan et al., 1994). DNA samples prepared from endodermal fraction were genotyped with respect to the Otx2 locusby PCR as described (Ang et al., 1996).

Mouse strainsThe Otx2neoallele was maintained on a CD1×129Sv/J background andanimals genotyped as described (Ang et al., 1996). The Lim1− allelewas maintained on a mixed (CD1×129Sv/J × C57BL/6) background.

847Sequential roles of Otx2 in brain patterning

itht,in

edfor

S

targeting vectorg3-hygro

E

B

N

X

(A)

S

1 kb

S

Pgk-hygro

N

E

EN

B

N

E E

BN E

N

Pgk-néo

N

Otx2neo alleleX H X E

X H X E

B

Otx2hygro allele

5’ probe

Pgk-hygro

NE

N

S

N AAXhoI XhoI

3’ probe

X H X

EAAAAA

Otx2 gene

XhoI XhoI

XhoI XhoI

XhoI XhoI

E

E

g1g3

E E

E

E E

E E

E

Fig. 1. Targeted disruption of the Otx2gene. (A) The open boxes represent the coding region and the solid boxes indicate the homeodomain ofOtx2. The DNA fragments contained in bacteriophage clones g3 and g1 are shown as dotted arrows above the genomic structure of the Otx2gene. The 5′ and 3′ probes used for Southern blot analysis are indicated. If homologous recombination occurs on the wild-type allele, digestionof genomic DNA with EcoRI shifts the restriction fragment identified by the 3′ flanking probe from 12.0 kb to 10.0 kb and digestion with NcoIshifts the restriction fragment identified by the 5′ flanking probe from 13.8 kb to 10.2 kb. Alternatively, if homologous recombination occurs onthe Otx2neoallele, digestion with EcoRI leads to a band shift from 13.4 kb to 10.0 kb using the 3′ flanking probe, while digestion with NcoIleads to a band shift from 12.5 kb to 10.2 kb using the 5′ flanking probe. B, BglII; E, EcoRI; N, NcoI; H, HindIII; S, StuI; X, XbaI. (B) Southernblot analysis of DNA from Otx2mutant ES cells. The sizes of the DNA bands are indicated in kb. The 5′ and 3′ probes detected restrictionfragments of predicted sizes for the wild-type, Otx2hygro and Otx2neoalleles. Abbreviations: WT, wild-type.

(B)

(A)

Mice carrying the ROSA26 gene trap insertion (Friedrich and Soria1991) were obtained from the Jackson Laboratories, Bar Harbou

Explant culturesThe ectoderm germ layer was separated from the mesendoderOtx2−/− and Lim1−/− embryos at E7.5 using tungsten needles, afincubating these embryos in 0.25% pancreatin, 0.5% trypsinphosphate-buffered saline (PBS) at 4°C for 10 minutes. The whectoderm, except for the posterior part corresponding to the primistreak, was used. Mesendoderm from the anterior half of wild-ty

no,r.

m ofter inoletivepe

embryos at E7.75 was isolated in a similar way and recombined wOtx2−/− or Lim1−/− ectoderm as described before (Ang and Rossan1993). Ectoderm from mutant embryos was also cultured alone bacteriological dishes. All explants were cultured for 2 days, and fixin 4% paraformaldehyde in PBS for 30 minutes and processed whole-mount in situ hybridization or antibody staining.

In situ hybridization and immunohistochemistryEmbryos were collected and fixed in 4% paraformaldehyde in PBfor 1 hour, rinsed in PBS plus 0.1% Tween 20 and stored at −20°C in

848

yle-

sly

.

as et

odyson

r axiorTosegeir

the

of

6).ral-

esed

ig.shsg

he

hatisthen,hen,

onor

M. Rhinn and others

Fig. 2.Lack of expression of anterior neural plate markers in Otx2−/−

embryos. Whole-mount RNA in situ hybridization of wild-type (A-C), and Otx2−/− (A′-C′) embryos. (A,A′) Six3expression at E7.75.Six3is expressed in the presumptive forebrain region in wild-typeembryos (A) but is missing in mutant embryos (A′). (B,B′) Fkh2expression at E7.75. In wild-type embryos, Fkh2 is expressed in theaxial mesoderm and the presumptive forebrain and midbrain (B). Inmutant embryos, Fkh2expression is restricted to axial mesodermcells at the distal tip (arrowhead in B′). (C,C′) Pax2 expression atE8.25. Pax2 is expressed in the mid-hindbrain region in wild-type(C) and mutant (C′) embryos, but the expression domain in mutantembryos is smaller in size presumably due to the absence of themidbrain. Anterior is to the left. Scale bar: 100 µm.

methanol. Chimeric embryos were identified morphologically or bβ-galactosidase staining of the entire embryo or the allantois. Whomount in situ hybridization was performed as described previou(Conlon and Herrmann, 1993). The following probes were used: Six3(Oliver et al., 1995), Fkh2(Kaestner et al., 1995), Pax2(Nornes et al.,1990), a 0.9 kb rat chordin cDNA (kindly provided by H. Wang),follistatin (Albano et al., 1994), a 1.3 kb noggin cDNA containingentire coding region (kindly provided by R. Harland and AMcMahon), Wnt1 (Fung et al., 1985), Sox1(Collignon et al., 1996),Hoxa2 and Hoxb1 (Wilkinson et al., 1989) and Mox1 (Candia et al.,1992). Whole-mount immunohistochemistry was performed described in Davis et al. (1991). The En-specific antiserum (Davisal., 1991) was used at a dilution of 1:1000. The secondary antibwas a peroxidase-coupled donkey anti-rabbit antibody (JackImmunoresearch) which was used at a dilution of 1:200.

RESULTS

Forebrain and midbrain are not induced in Otx2−/−

embryosWe have previously reported that embryos homozygous fomutation in Otx2 consisting of a deletion of the homeoboregion are missing the forebrain, midbrain and anterhindbrain at early somite stages (Ang et al., 1996). determine if this defect is due to a lack of formation of thetissues, which are normally induced by the underlyinmesendoderm, or to a loss of the tissues shortly after thformation, we studied the expression of early markers of anterior neural plate in Otx2−/− embryos. Six3 is a homeoboxgene that is first expressed only a few hours after formationthe anterior neural plate at E7.75 (headfold stage), in presumptive forebrain region (Oliver et al., 1995). Six3was notexpressed in Otx2−/− embryos of E7.5-8.25 (late-streak to earlsomite stages) (Fig. 2A,A′ and data not shown). Fkh2, awinged-helix gene expressed in the presumptive forebramidbrain and anterior axial mesoderm (Kaestner et al., 199was not expressed in the neural plate, but was observed in amesendoderm cells in the distal tip of Otx2−/− embryos at E7.75(headfold stage) (Fig. 2B,B′). The lack of early markers ofpresumptive forebrain and midbrain strongly suggests a failof induction of these tissues in Otx2−/− embryos rather than aloss of the tissues subsequently to their formation.

We also examined whether anterior hindbrain, and particular presumptive rhombomere 1, was initially formed Otx2−/− embryos, using Pax2 (Rowitch and McMahon, 1995)as a marker of this region. Pax2 was expressed in a broaddomain corresponding to the midbrain and anterior hindbrin wild-type embryos at E8.25 (Fig. 2C). Pax2 was stillexpressed in Otx2−/− embryos, albeit in a reduced domain (Fig2C′). The reduced domain of Pax2 expression is most likelydue to expression in the anterior hindbrain of Otx2−/− embryos.Thus, in contrast to the lack of formation of forebrain anmidbrain, anterior hindbrain tissue corresponding rhombomere 1 appears to be formed initially but subsequently deleted in Otx2−/− embryos at E8.5 (Ang et al.,1996).

noggin, chordin and follistatin are expressed inOtx2−/− embryosThe lack of formation of the anterior neural tissue in Otx2−/−

embryos could result from a loss of neural-inducing molecu

the

y

in,5),xial

ure

inin

ain

.

dtois

les

in the organizer, or node (Lemaire and Kodjabachian, 199We therefore studied the expression of three candidate neuinducing genes, noggin (Smith and Harland, 1992), chordin(Sasai et al., 1994) and follistatin (Hemmati-Brivanlou et al.,1994), in mutant embryos. Although no organised nodstructure could be recognised, the three genes were expresin the distal tip of Otx2−/− embryos, i.e. in the regioncorresponding to the node in wild-type embryos (Fig. 3A′-C′).Nogginand chordin, which are normally also expressed in theaxial mesoderm at the anterior end of wild-type embryos (F3A,B), were missing in the anterior midline of mutant embryo(Fig. 3A′,B′). Follistatin was also expressed in the entire lengtof the primitive streak of both wild-type and mutant embryo(Fig. 3C,C′). These results confirm previous studies showindefects in the anterior axial mesoderm of Otx2−/− mutantembryos (Ang et al., 1996). They also demonstrate that tfailure to induce forebrain and midbrain in Otx2−/− embryos isnot due to a loss of expression of noggin, chordinor follistatinin the node.

Generation of strong Otx2−/−↔+/+ chimerasEmbryological studies in vertebrates have demonstrated tthe formation and regionalization of the neural tube dependent on interactions between the ectoderm and underlying axial mesendoderm (Lemaire and Kodjabachia1996). In mouse, recent studies have also implicated tvisceral endoderm in this process (Thomas and Beddingto1996; Varlet et al., 1997). Since Otx2 is expressed in all germlayers at the anterior end of the embryo during gastrulati(Simeone et al., 1993; Ang et al., 1994), the lack of anteri

849Sequential roles of Otx2 in brain patterning

e or

uire

ithrsed

rictsto

ig.ge,bleinicanldsntin

at

edse

Fig. 3. Expression of signalling molecules in Otx2−/− embryos.Whole-mount RNA in situ hybridization of wild-type (A-C) andOtx2−/− embryos (A′-C′) at E7.75. (A,A′,B,B′) noggin(A) andchordin(B) are expressed in the axial mesoderm and node in normalembryos. In mutant embryos, noggin(A′) and chordin(B′) areexpressed in the disorganised node but expression in the axialmesoderm does not extend as far anteriorly as in wild-type embryos.(arrows in A′,B′) Follistatin is expressed in the primitive streak ofwild-type (C) and Otx2−/− mutant embryos (C′). Anterior is to theleft. Scale bar: 100 µm.

Fig. 4.Morphological analysis of Otx2−/−↔+/+ chimeric embryos.(A,D) Wild-type, (B,E) Otx2−/−↔+/+ and (C, F) Otx2−/− embryos.(A-C) 0- to 4-somite stage. Otx2−/−↔+/+ chimeric embryos (B)show normal development of anterior neural folds (arrowheads),compared to wild-type embryos (A). This tissue is missing at theanterior end of Otx2−/− embryos (C). (D-F) 6- to 8-somite stage.Otx2−/−↔+/+ chimeric embryos (E) show better development of therostral brain than Otx2−/− embryos (F). However, the forebrain inchimeric embryos appeared less well developed than in wild-typeembryos (arrows in E and D, respectively). Anterior is to the top,except for C where anterior is to the left. WT, wild-type. Scale bar:100 µm.

neural plate in Otx2−/− embryos could be due to defects in thinducing anterior axial mesendoderm or visceral endodermin the responding ectoderm. To identify the tissues that reqOtx2activity for anterior neural plate formation, we generatchimeric mouse embryos composed of Otx2−/− and wild-typecells, referred to as Otx2−/−↔+/+ chimeras. These chimerawere produced by injecting three different Otx2−/− ES celllines, called TX3, TX62 and TX112, into morula-stagembryos generated by crossing Rosa26homozygous maleswith CD1 females. The Rosa26transgenic mouse line carriea lacZ gene trap insertion that results in widespread lacZexpression during embryonic development (Friedrich et 1991). Thus, expression of lacZ was used to distinguish cellsderived from wild-type embryos (lacZ-positive) from Otx2−/−

ES-derived (lacZ-negative) cells. Reimplanted chimeriembryos were harvested between E8.5 and E9.5 and the of chimerism was analysed by immunohistochemical stainfor β-galactosidase activity. We have classified the chimerastrong, moderate or weak depending on whether they conmore than 90%, 50% or 20%, respectively, of Otx2−/− cells. Inthis paper, we will report only our studies of the stronchimeras. In these chimeras, the embryo is almost entimade up of Otx2−/− cells, while the VYS contains exclusivelywild-type cells (see below). We were thus able to use thchimeras to address the function of Otx2 in the visceralendoderm.

Rescue of anterior neural plate in strong chimerasWe harvested a total of 49 strong Otx2−/−↔+/+ chimerasbetween E8.5 and E9.5. Strong chimeras (referred to belowchimeric embryos or chimeras) were developmentally delay

ed

s

e

s

al.,

clevelings astain

grely

ese

ased,

and contained 0 to 14 somites (Figs 4B,E, 5C), compared wnon-chimeric embryos or weak chimeras in the same litte(data not shown) and with control chimeric embryos generatby injection of +/+ or Otx2+/− ES cells (Fig. 5A,B), that had12 to 40 somites. The delayed development of the chimeembryos suggests that Otx2regulates the growth of the epiblasduring gastrulation. Interestingly, all the chimeric embryowith less than 10 somites contained tissue anterior rhombomere 2 (Fig. 4B,E), which was missing in Otx2−/−

mutant embryos obtained from heterozygous intercrosses (F4C,F). In several chimeric embryos at the 0- to 4-somite stathe rescued anterior neural plate appeared indistinguishafrom that of wild-type embryos (compare panels B and A Fig. 4). In contrast, the anterior neural plate in chimerembryos with 6-8 somites appeared less well developed thin wild-type embryos at the same stage (Fig. 4D,E) and it couno longer be identified morphologically in chimeric embryowith 12-14 somites (data not shown). Thus, the developmeof the rostral neural plate appears to be initially rescued chimeric embryos, but this tissue is progressively deletedlater stages.

Distribution of mutant and wild-type cells in strongchimerasTen chimeras containing 0-6 somites were analyshistologically in transverse sections. The neural tube of the

850

m.s

riorhisby

.5

dseederal

in

tral

etdediorands0-4r

was

M. Rhinn and others

Fig. 5.Control chimeras. (A-C) Comparison of the distribution of embryonic cells and ES cells in strong chimeric embryos, generated byinjection of +/+ (A), Otx2+/− (B) and Otx2−/− (C) ES cells into wild-type RosA26/+ embryos at E9.5 and stained for β-galactosidase activity.lacZ-positive, wild-type cells derived from the embryo are located exclusively in the gut (arrows) and VYS endoderm (arrowheads) in thesedifferent chimeras (A-C), irrespective of the genotype of the ES cells. (D) Reciprocal type of chimeras generated by injection of wild-typeROSA26/+ ES cells into a Otx2−/− morula at E8.5. lacZ-positive, wild-type ES cells have not colonized the VYS endoderm which is entirelyOtx2−/− (see text), but have extensively colonized the embryo and mesoderm of the VYS. Nevertheless, the Otx2mutant phenotype is notrescued, thus demonstrating that this phenotype originates from a requirement for Otx2function in the visceral endoderm. Anterior is to the top.Scale bar: 100 µm.

chimeras was entirely composed of Otx2−/− cells (Figs 5C, 6D-I). The tissue rostral to the hindbrain had the characterimorphology of a neuroepithelium, suggesting that it hadneural identity (Fig. 6D,G). Otx2−/− cells were also observedin the prechordal plate (Fig. 6D) and anterior notochord (F6E,H). Thus, Otx2−/− cells can contribute to all the tissues ostrong chimeric embryos, including the rostral braprechordal plate and anterior notochord, which are missingOtx2−/− embryos.

In chimeric embryos, wild-type cells were found exclusivein the endoderm of the VYS and of the gut along the entireP axis (blue cells, Fig. 6D-I). In two of the ten chimeranalysed, wild-type cells were in addition found in notochocells (data not shown). In the gut endoderm, wild-type cewere intermingled with mutant cells and contributed from to 50% of this germ layer (Fig. 6E-I). In contrast, the VYendoderm contained exclusively wild-type cells. The abseof ES-derived cells in visceral endoderm derivatives was surprising, since it has previously been shown that ES ccontribute very poorly to this tissue (Beddington anRobertson, 1989; Varlet et al., 1997). However, ES cells hnot been shown to contribute less efficiently than embryocells to the gut endoderm.

We therefore examined whether the presence of wild-tycells in the gut endoderm of ES cell chimeras correlated wthe mutation of Otx2 in ES cells. Two +/+ ES cell lines, R1(Nagy et al., 1993) and H1 (A. Dierich, unpublished), and tOtx2+/− ES cell lines, TX106 and TX111, were injected inmorula-stage embryos. In chimeras generated with these +/+and Otx2+/− ES cell lines, referred to as +/+ ↔+/+ and Otx2+/−

↔+/+ , respectively, lacZ-positive embryo-derived cells werealso found in the endoderm of the VYS and of the gut (F5A,B, 6A-C), demonstrating that the presence of wild-tycells in the gut in chimeric embryos is not due to the Otx2mutation.

stic a

ig.f

in, in

ly A-asrdlls

10Sncenotellsd

avenic

peith

woto

igspe

Wild-type visceral endoderm rescues anterior neuralplate in chimeric embryosHistological analyses of Otx2−/−↔+/+ chimeras demonstratedthat wild-type cells were always found in the VYS endoderIn addition, wild-type cells were found exclusively in thitissue in 8/10 chimeras. These results suggest that Otx2 isrequired in the visceral endoderm for the rescue of anteneural plate development in chimeric embryos. To confirm tresult, we generated the reciprocal type of chimeras, injecting wild-type ROSA26/+ (lacZ-positive) ES cells intoOtx2−/− embryos (referred to as +/+ ↔Otx2−/− chimeras). All6 +/+ ↔Otx2−/− strong chimeric embryos examined at E8exhibited the characteristic Otx2 homozygous mutantphenotype (Fig. 5D), failing to develop a forebrain anmidbrain. PCR analysis of the endoderm of the VYS of thechimeras confirmed the mutant genotype of the injectembryos and also showed that in four cases, the viscendoderm contained exclusively Otx2−/− mutant cells (data notshown). Altogether, these results demonstrate that Otx2 isrequired in the visceral endoderm for induction of forebraand midbrain.

Lack of regional specification in the rescuedanterior brain of chimeric embryosTo first confirm the neural nature of the rescued tissue rosto the hindbrain of Otx2−/−↔+/+ chimeras, we showed that itexpressed Sox1, an early general neural marker (Collignon al., 1996, Fig. 7A,A′). We next determined whether the rescueOtx2−/− neural tissue was properly regionalised. We analysthe expression of genes marking different anterior-posterpositions along the neural tube in chimeras between E8.5 E9.5. Six3, an early forebrain marker (Oliver et al., 1995) waexpressed in a reduced domain in chimeras containing somites (Fig. 7B,B′). Its expression, however, was no longedetected in chimeras at the 6- to 8-somite stage whereas it

851Sequential roles of Otx2 in brain patterning

ic-

e

inlly

Fig. 6. Histological analysis of strong Otx2−/−↔+/+ chimeric embryos. (A-C) +/+↔+/+ control chimera; (D-I). Otx2−/−↔+/+ chimeras.(A,D,G) Transverse sections at the head level, (B,E,H) at the heart level and (C,F,I) at the spinal cord level. (D-I) The anterior neural tube inOtx2−/−↔+/+ chimeras contains only Otx2−/− (pink) cells and presents a morphology characteristic of neuroepithelium. Otx2−/− cells alsocontribute to the prechordal plate (open arrow in Fig. 6D) and the anterior notochord (open arrowheads). In contrast, wild-type (blue) cells arelocalised exclusively in the endoderm of the VYS (arrowheads in B-I) and of the gut (arrows in B,C, E-I) in both +/+ ↔+/+ control chimeras(B,C) and Otx2−/−↔+/+ chimeras (D-I). In the gut endoderm, the ES-derived cells are intermingled with wild-type cells, whereas the VYS ismade up exclusively of wild-type cells (C-I). Scale bar: 100 µm.

normally expressed in the ventral diencephalon and opvesicles of control embryos (7C,C′). In contrast, expression ofHesx1/Rpx (Thomas and Bedddington, 1996; Hermesz et a1996), another early forebrain-specific gene, was neobserved in chimeric embryos, even between the eaheadfold and 4-somite stage (Fig. 7D,D′,E,E′ and data notshown). Therefore, different forebrain genes show differerequirement for Otx2 function. Otx2 is required for theinitiation of Hesx1/Rpx expression, and for the maintenance Six3expression.

tic

l.,verrly

nt

of

Expression of the regional markers for the mesencephalmetencephalic (mes-met) region, Pax2 (Nornes et al., 1990),En (Davis et al., 1991) and Wnt1(Rowitch et al., 1995), wasalso analysed in Otx2−/−↔+/+ chimeric embryos. At the 0-to 4-somite stage, Pax2expression in the mes-met region wasexpanded rostrally compared to its expression in wild-typembryos (Fig. 8A,A′). At the 4-somite stage, En expressionin chimeric embryos was weaker and more patchy than wild-type embryos, and also appeared expanded rostra(Fig. 8C,C′). By the 6- to 8-somite stage, Pax2 and En

852

2-

hatnotrs

oth

ndtwofor

ofshthee

M. Rhinn and others

Fig. 7. Expression of the neural markers, Sox1, Six3, Hoxa2, Hesx1/Rpx, andHoxb1in Otx2−/−↔+/+ chimeras. Whole-mount RNA in situhybridization analysis of wild-type embryos (A-F) and Otx2−/−↔+/+ chimeras (A′-F′). (A,A′) Sox1 expression at the 6-8-somite stage. Sox1issimilarly expressed in the neural tube of wild-type (A) and chimeric embryos (A′). (B,B′) Six3expression at the 0- to 4-somite stage. Six3isexpressed in a reduced domain in the anterior neural plate (presumptive forebrain) of chimeric embryos (B′), compared to its expression inwild-type embryos (B). (C,C′) Six3 and Hoxa2expression at the 6- to 8-somite stage. Six3 is no longer detected in chimeric embryos (C), whileit is still expressed in the forebrain and optic vesicles of wild-type embryos (C′). In contrast, Hoxa2is similarly expressed in rhombomere 3(arrow) in wild-type (C) and chimeric (C′) embryos. (D,D′,E,E′) Hesx1/Rpx expression at the 0- to 4 (D) and 6- to 8 (E)-somite stages in theforebrain was not detected in chimeric embryos (D′,E′). (F,F′) Hoxb1is expressed in rhombomere 4 and the posterior spinal cord in both wild-type (F) and chimeric embryos at the 6- to 8-somite stage (F′). Anterior is to the top. WT, wild-type. Scale bar: 100 µm.

expression domains were not maintained except for a vsmall domain at the rostralmost tip of the chimeric embry(Fig. 8B,B′,D,D′). Wnt1 is normally expressed at E8.5 in dorsal midline stripe along the neural tube and, intransverse band, in the caudal mesencephalon. The doexpression domain of Wnt1was present, but its expression ithe ventral mesencephalon was not detected in chimembryos containing 0 to 8 somites (Fig. 8E,E′,F,F′).Therefore, as for forebrain genes, the expression of mes-genes is differentially affected by the loss of Otx2 function.Otx2 is required for the initiation of Wnt1 expression inventrocaudal mesencephalon and for the maintenancePax2and En expression in the mes-met region. In additioEn, which is normally expressed in wild-type embryos by t1-somite stage (Davis et al., 1991), was not detectedchimeric embryos with 1-3 somites (data not showindicating a delay in the initiation of En expression in thabsence of Otx2.

The expression of hindbrain-specific genes was aexamined in chimeric embryos. Expression of the genes Hoxa2and Hoxb1 (Wilkinson et al., 1989) in rhombomeres 3 (Fig

eryosa arsal

neric

met

ofn,he in

n),e

lso

.

7C,C′) and 4 (Fig. 7F,F′) respectively, were similar in wild-typeand chimeric embryos at the 6- to 8-somite stage. At the 1somite stage, expression of Hoxa2 in rhombomere 3 wasobserved close to the rostral end of chimeras, indicating ttissues anterior to rhombomere 3 were deleted (data shown). Therefore, expression of posterior neural markeoccurred normally in the absence of Otx2expression.

Otx2 is required in the neurectoderm for Engrailedexpression Rescued anterior neurectoderm and axial mesoderm bcontained exclusively mutant cells in most Otx2−/−↔+/+chimeric embryos. The defects in expression of forebrain ames-met markers in these chimeras could therefore have causes: it could be due to an autonomous requirement Otx2 in the neurectoderm, or to a requirement for Otx2 in theaxial mesoderm for the induction or maintenance expression of regional markers in the brain. To distinguibetween these two hypotheses and specifically to study function of Otx2 in the neurectoderm, we performed culturexperiments involving the recombination of ectoderm from

853Sequential roles of Otx2 in brain patterning

ft

toed

rrm.isrm

ilar.

f 6ts

Fig. 8.Expression of mes-met markers in Otx2−/−↔+/+ chimeras. Labelling of wild-type embryos (A-F) and Otx2−/−↔+/+ chimeras (A′-F′).(A,A ′) Pax2 expression at the 0- to 4-somite stage. Pax2RNA expression in the presumptive mes-met region (brackets) is enlarged towards therostral end in chimeric embryos (A′), compared to its expression in wild-type embryos. (B,B′) Pax2RNA expression at the 6- to 8-somite stage.Pax2is expressed in a small domain at the rostral end (arrow) of chimeric embryos, whereas it is expressed in two separate domains in theforebrain and mes-met region of wild-type embryos. (C,C′) En protein expression at the 4-somite stage. En is expressed in the mes-met regionof wild-type embryos (bracket in C), and has a weaker and more patchy expression domain, which is expanded rostrally in chimeric embryos(bracket in C′). (D,D′) En protein expression at the 6- to 8-somite stage. En is expressed in the mes-met of wild-type embryos (D), but only atthe rostralmost tip of the neural tube of chimeric embryos (arrow in D′). (E,E′) Wnt1and Mox1expression at the 0- to 4-somite stage. Wnt1isexpressed in a weak and broad domain in the mesencephalon of wild-type embryos (brackets in E), but not in a chimeric embryo with threesomites expressing Mox1(small arrows in E′). (F,F′). Wnt1expression at the 6- to 8-somite stage. Wnt1 is expressed in the dorsal neural tubebut not in a transverse band in the posterior mesencephalon of chimeric embryos (F′), whereas it is expressed in both these regions in wild-typeembryos (F). Anterior is to the top. WT, wild-type. Scale bar: 100 µm.

Otx2−/− embryos with axial mesendoderm from wild-typembryos. We have previously shown in similar explarecombination experiments that anterior mesendoderm frwild-type embryos can induce expression of the mes-mmarker En in posterior lateral ectoderm isolated from wiltype embryos (Ang and Rossant, 1993). Thus, posteectoderm cells, whose normal fate is spinal cord and surfectoderm, can be instructed to adopt a mes-met fateanterior mesendoderm. Using this assay, we asked wheOtx2 is required in the neuroectoderm for the expressionEn. The ectoderm germ layer from Otx2−/− embryos at E7.5was recombined with the anterior mesendoderm from E7wild-type embryos. These explants were cultured for 2 daand then assayed for expression of En proteins by whmount antibody staining.

All ten explants of Otx2−/− ectoderm recombined withanterior mesendoderm from wild-type embryos failed

entomet

d-riorace byther of

.75ys

ole-

to

express En after culture (Fig. 9B). Six control explants oOtx2−/− ectoderm cultured alone for 2 days also did noexpress En (Fig. 9A). Lack of En expression was not duea failure of Otx2−/− ectoderm cells to adopt a neural fate, sincOtx2−/− ectoderm isolated from E7.5 embryos and culturefor 2 days expressed the general neural marker,Sox1(Fig.9C). As a control experiment for the efficiency of germ layeinteractions in this culture system, we recombined ectodefrom Lim1−/− embryos with wild-type anterior mesendodermLim1 is a LIM and homeobox domain-containing gene that expressed in the axial mesendoderm but not in the ectodeor the neural plate (Barnes et al., 1994). Lim1−/− embryospresent a loss of brain tissue rostral to rhombomere 3, simto that of Otx2−/− embryos (Shawlot and Behringer, 1995)When we recombined Lim1−/− ectoderm with wild-typemesendoderm, En expression was rescued in 3 out omutant ectoderm explants (Fig. 9D). These resul

854

5.we

l

id-f

,h

f

f

s

8s,

se

.

toly

fddty.

),

ad

nal

ng

l

in

M. Rhinn and others

Fig. 9. Expression of En and Sox1in explants after culture for 48hours. (A) Ectoderm explants from E7.5 Otx2−/− embryos fail toexpress En. (B) Ectoderm explants from E7.5 Otx2−/− embryosrecombined with anterior mesendoderm from E7.75 wild-typeembryos do not express En. (C) Ectoderm explants from E7.5Otx2−/− embryos showing expression of the general neural markeSox1. (D) Ectoderm explants from E7.5 Lim1−/− embryos,recombined with anterior mesendoderm from E7.75 wild-typeembryos, express En. Scale bar: 100 µm.

demonstrate that Lim1 is not required in the neurectoderm foEn expression, and that the loss of the mes-met regionLim1−/− embryos must be due to a requirement for Lim1 inaxial mesendoderm or visceral endoderm, two tissues normally expresses Lim1 during gastrulation (Barnes et al.1994; our unpublished results). In contrast, Otx2 function isrequired in the neurectoderm for expression of En proteiSpecifically, since the explants were analysed for expression 2 days after culture at a stage correspondinapproximately E9.0-E9.5, the recombination assay measuthe requirement for Otx2 in the maintenance rather than thinitiation of En expression, which occurs in mouse embryat E8.0. Thus, Otx2 is required autonomously in theneurectoderm for the maintenance of En expression,agreement with the results obtained from chimeric studie

DISCUSSION

In this paper, we show that the absence of rostral brainOtx2−/− embryos is due to a failure in the induction of thanterior neural plate at early stages of gastrulation. In strOtx2−/−↔+/+ chimeras, induction of forebrain and midbraiis rescued by the presence of wild-type cells in the visceendoderm. Altogether, these results demonstrate that Otx2 isrequired in the visceral endoderm for induction of the rostbrain. Using chimera studies and an explant recombinatassay, we then demonstrate that Otx2−/− ectodermal cells canacquire some but not all the characteristics of anterneuroectodermal cells. Therefore, Otx2has a second functionin the neuroectoderm for the specification of the forebrain ames-met regions of the neural tube.

r in

that,

ns.Eng tores

eos

ins.

ine

ongnral

ralion

ior

nd

Otx2 is required in the visceral endoderm forinduction of forebrain and midbrainIn previous studies, we and others demonstrated that Otx2−/−

embryos lack neural tissue rostral to rhombomere 3 at E8.Using early markers of the anterior neural plate, we now shothat the presumptive forebrain and midbrain do not form whila presumptive rhombomere 1 expressing Pax2 appears to begenerated in Otx2−/− embryos at E7.75. Thus, the loss of neuratissue at this early stage correlates with the domain of Otx2expression in the neural tube, which ends precisely at the mhindbrain boundary (Millet et al., 1996). The absence oexpression of Six3and Fkh2 in Otx2−/− embryos indicates thatthe defect in anterior neurectoderm occurs very earlypresumably during the phase of induction of this tissue througectoderm-mesendoderm interactions.

Our chimeric studies provide strong evidence for a role oOtx2 in the induction of forebrain and midbrain tissues. Inchimeras, a general inability of ES cells to contribute tovisceral endoderm leads to the preferential localisation oembryo-derived wild-type cells in this tissue along the wholeanterior-posterior axis of the embryos. In strong Otx2−/− ↔+/+chimeras (made of more than 90% Otx2−/− cells), the neuraltissue rostral to the hindbrain always developed and wacomposed entirely of Otx2−/− cells, whereas wild-type cellswere found exclusively in the visceral and gut endoderm in out of 10 chimeras. Furthermore, the reciprocal chimeraobtained by injecting +/+ cells into Otx2−/− embryos andhaving a visceral endoderm containing only Otx2−/− mutantcells, showed no rescue of rostral brain development. Theresults demonstrate that Otx2 function is required in thevisceral endoderm for development of forebrain and midbrainWhether wild-type cells in the gut are also involved in therescue remains to be determined, since it is not possible generate chimeric embryos whose gut endoderm is entiremade up of Otx2−/− cells to test this hypothesis.

Our experiments provide compelling evidence for a role oOtx2 in an interaction between visceral endoderm anectoderm, necessary for the induction of forebrain anmidbrain in mouse (Thomas and Beddington, 1996; Varlet eal., 1997). The nature of the signalling molecules regulated bOtx2and involved in this interaction remains to be determinedTwo candidates are nodal, a TGFβ-related growth factorinvolved in anterior neural plate patterning (Varlet et al., 1997and the secreted molecule cerberus, which is expressed inanterior endoderm and has the property to induce ectopic hestructures when microinjected into ventral regions of Xenopusembryos (Bouwmeester et al., 1996). A mouse cerberushomologue (mCer-1) has recently been isolated and present aabnormal and reduced expression domain in the viscerendoderm of Otx2−/− embryos (Biben et al., 1997).

Otx2 is required in the neuroectoderm for itsregional specificationTo study Otx2 function in the neuroectoderm, we have furthercharacterised the anterior neural tissue that is rescued in strochimeric embryos. This tissue, entirely made of Otx2−/− cells,expresses Sox-1 and acquires some measure of regionapatterning. Expression of Six3 is initiated, albeit in a reduceddomain, but is not maintained. Expression of Hesx1/Rpxis notinitiated in the forebrain. These results indicate that a forebratissue is formed in absence of Otx2, but that Otx2 is required

r,

855Sequential roles of Otx2 in brain patterning

asr,l

rm

alaffsain

sy

.-L.téche

ein

al

oic

d

n

ndis.

ne

rior

for its regional specification as early as the 0- to 4-somite staIn addition, the expression domains of mes-met markers, sas Pax2 and En, are expanded rostrally in chimeric embryosthis stage. This result, together with the reduced domain of Six3expression, suggest that part of the presumptive forebrain adopted a more posterior fate and that Otx2 may be requiredto prevent forebrain cells adopting a mes-met fate. By theto 8-somite stage, a neural tissue expressing forebrain maris missing altogether, indicating that the forebrain has eitbeen eliminated or completely transformed to a more postefate. Thus, Otx2 is also required after the initial stage oinduction and specification, for maintaining regional identiin the forebrain and/or maintaining the tissue itself.

Pax2and En are initially expressed in the mes-met regionOtx2 chimeras, suggesting that establishment of this territooccurs in the absence of Otx2 function. However, Otx2 isrequired within this territory for the distinction betweemesencephalic and metencephalic fates, since in its abseexpression of Wnt1, a gene required for development of thmesencephalon (McMahon et al., 1992), is not initiated. addition, Pax2 and En expression are almost completely loby the 6- to 8-somite stage, indicating that Otx2 is also requiredfor the maintenance of the mes-met region. Loss of expression could in part be caused by the loss of Wnt1expression, since Wnt1 is required for the maintenance of Eexpression (McMahon et al., 1992; Danielian and McMaho1996). Interestingly, the regulation of En and Wnt1 by Otx2appears to be evolutionarily conserved since en and wg aretargets of the Otx homolog, Otd, in flies (Cohen and Jurgens1991).

Analysis of regional markers demonstrate important rofor Otx2 in establishing regional identities in both forebraiand midbrain. One mechanism by which Otx2could establishdistinct fates in these two regions could be by regulating geexpression in combination with other regionally expressgenes. In the forebrain, Otx2would interact with a coactivatorto initiate expression of Hesx1/Rpx. In the mes-met region,Otx2would interact with another coactivator, possibly Pax2, toinitiate Wnt1 expression and consequently specifmesencephalic fates. This hypothesis raises the possibility Hesx1/Rpx and Wnt1 are direct targets of Otx2. By studyingHesx1/RpxandWnt1 expression in moderate Otx2−/−↔+/+chimeras, in which the neuroectoderm contains a mosaicmutant and wild-type cells, we should be able to determineOtx2 is required cell-autonomously or non-autonomously fthe expression of these genes.

In our discussion so far, we have assumed that Otx2functions in the neuroectoderm for establishing forebrain amidbrain identity. However, we cannot exclude the possibilthat some of the neural plate patterning defects of the chimeare due to a requirement for Otx2 in the prechordal plate. Wehave nevertheless addressed the specific requirement for Otx2function in the neuroectoderm for the expression of oregional marker, using explant-recombination experimenThese experiments demonstrate that Otx2 is required in theneurectoderm for the maintenance of En expression.

In strong Otx2−/−↔+/+ chimeras, the prechordal plate ananterior notochord contained exclusively Otx2−/− cells andnevertheless developed normally (expression of HNF-3β inthese tissues was normal, data not shown). Therefore, Otx2 isnot required cell-autonomously for the formation of the

ge.uch at

has

6-kersherriorfty

ofry

nnce,eInst

En

nn,

,

lesn

need

ythat

of if

or

ndityras

nets.

d

se

tissues. In addition, Otx2is also not required in axial mesodermand/or node for expression of signalling molecules such chordin, noggin and follistatin. Our experiments, howevecannot rule out that Otx2has also a function in the prechordaplate for regionalization of the neural tube, since Otx2activitywas missing in both the neuroepithelium and axial mesodeof strong Otx2−/−↔+/+ chimeras. Alternative approaches,involving the specific knock-out of Otx2 function in differentgerm layers, will be necessary to study if Otx2 also has afunction in the prechordal plate.

We thank Drs Rosa Beddington and François Guillemot for criticreading of the manuscript and Valerie Meyer and the transgenic stfor excellent technical assistance. We are grateful to Drs RoBeddington, Peter Gruss, Richard Harland, Klaus Kaestner, RobLovell-Badge, Andy McMahon, Hai Wang, Marion Wassef, DavidWilkinson and Chris Wright for in situ probes. The En antiserum wagenerously provided by Dr Alex Joyner. This work was supported ba research grant from the Human Frontier Science Program to SA. and R. R. B. and by funds from the Institute National de la Sanet de la Recherche Médicale, the Centre National de la RecherScientifique, and the Centre Hospitalier Universitaire Régional.

REFERENCES

Acampora, D., Mazan, S., Lallemand, Y., Avantaggiato, V., Maury, M.,Simeone, A. and Brulet, P. (1995). Forebrain and midbrain regions aredeleted in Otx2−/− mutants due to a defective anterior neuroectodermspecification during gastrulation. Development121, 3279-3290.

Albano, R. M., Arkell, R., Beddington, R. S. and Smith, J. C.(1994).Expression of inhibin subunits and follistatin during postimplantation mousdevelopment: decidual expression of activin and expression of follistatin primitive streak, somites and hindbrain. Development120, 803-813.

Ang, S.-L. and Rossant, J.(1993). Anterior mesendoderm induces mouseEngrailed genes in explant cultures. Development118, 139-149.

Ang, S.-L., Conlon, R. A., Jin, O. and Rossant, J. (1994). Positive andnegative signals from mesoderm regulates the expression of mouse Otx2 inectoderm explants. Development 120, 2979-2989.

Ang, S.-L., Jin, O., Rhinn, M., Daigle, N., Stevenson, L. and Rossant, J.(1996). A targeted mouse Otx2 mutation leads to severe defects ingastrulation and formation of axial mesoderm and to deletion of rostrbrain. Development122, 243-252.

Bally-Cuif, L., Gusilano, M., Broccoli, V. and Boncinelli, E.(1995). c-otx2is expressed in two different phases of gastrulation and is sensitive to retinacid treatment in chick embryo. Mech. Dev.49, 49-63.

Barnes, J. D., Crosby, J. L., Jones, C. M., Wright, C. V. E. and Hogan, B.L. M. (1994). Embryonic expression of Lim-1, the mouse homolog ofXenopus XLim-1, suggests a role in lateral mesoderm differentiation anneurogenesis. Dev. Biol.161, 168-178.

Beddington, R. S. P. and Robertson, E. J.(1989). An assessment of thedevelopmental potential of embryonic stem cells in the midgestatioembryo. Development 105, 733-737.

Beddington, R. S. P, Morgenstein, J., Land, H. and Hogan, A. (1989). Anin situ transgenic enzyme marker for the midgestation mouse embryo athe visualization of the inner cell mass clones during early organogenesDevelopment106, 37-46.

Biben, C., Stanley, E., Fabri, L., Kotecha, K., Rhinn, M., Drinkwater, K.,Wang, C., Nash, A., Hilton, D., Ang, S.-L., Mohun, T. and Harvey, R.P. (1997). A murine cerberushomologue (mCer-1) expressed in anteriorprimitive endoderm and embryonic mesendoderm during gastrulation. Dev.Biol., in press.

Blitz, I. C. and Cho K. W. Y. (1995). Anterior neurectoderm is progressivelyinduced during gastrulation: the role of the Xenopus homeobox georthodenticle. Development121, 993-1004.

Bouwmeester, T., Kim, S.-H., Sasai, Y., Lu, B. and De Robertis, E. M.(1996). Cerberus is a head-inducing secreted factor expressed in the anteendoderm of Spemann’s organizer.Nature382, 595-601.

Candia, A. H., Hu, J., Crosby, J., Lalley, P. A., Noden, D., Nadeau, J. H.and Wright, C. V. E. (1992). Mox-1 and Mox-2 define a novel homeobox

856

.

is

dy

l

s

s

o.

M. Rhinn and others

gene subfamily and are differentially expressed during early mesoderpatterning in mouse embryos. Development 116, 1123-1136.

Chen, W. S., Manova, K., Weinstein, D. C., Duncan, S. A., Plump, A. S.,Prezioso, V. R., Bachvarova, R. F. and Darnell, J. E., Jr.(1994).Disruption of the HNF-4gene, expressed in visceral endoderm, leads to cdeath in embryonic ectoderm and impaired gastrulation of mouse embrGenes Dev.8, 2466-2477.

Cohen, S. and Jurgens, G. (1991). Drosophila headlines. Trends Genet.7,267-272.

Collignon, J., Sockanathan, S., Hacker, A., Cohen-Tannoudji, M., Norris,D., Rastan, S., Stevanovic, M., Goodfellow, P. N. and Lovell-Badge, R.(1996). A comparison of properties of Sox-3 with Sryand two related genes,Sox-1and Sox-2. Development122, 509-520.

Conlon, R. A. and Herrmann, B. G. (1993). Detection of messenger RNAby in situ hybridization to postimplantation embryo whole mounts. MethodsEnzymol. 225, 373-383.

Danielian P. S. and McMahon A. P.(1996). Engrailed-1as a target of theWnt-1signalling pathway in vertebrate midbrain development. Nature383,332-334.

Davis, C. A. and Joyner, A. L. (1988). Expression patterns of the homeoboxcontaining genes En-1and En-2and the proto-oncogene int-1 diverge duringmouse development. Genes Dev.2, 1736-1744.

Davis, C. A., Holmyard, D. P., Millen, K. J. and Joyner, A. L.(1991).Examining pattern formation in mouse, chicken and frog embryos with En-specific antiserum. Development111, 287-298.

Dierich, A. and Dollé, P. (1997). Gene targeting in embryonic stem cells. IMethods in Developmental Toxicology and Biology, pp111-123. Oxford:Blackwell Science.

Duncan, S. A., Nagy, N. and Chan, W.(1997). Murine gastrulation requiresHNF-4 regulated gene expression in the visceral endoderm: tetraploid resof HNF-4−/− embryos. Development124, 279-287.

Finkelstein, R. and Perrimon, N. (1991). The molecular genetics of headdevelopment in Drosophila melanogaster. Development112, 899-912.

Friedrich, G. and Soriano, P.(1991). Promoter traps in embryonic stem cellsa genetic screen to identify and mutate developmental genes in mice. GenesDev.5, 1513-1523.

Fung, Y. K., Shackleford, G. M., Brown, A. M. and Sanders, G. S. (1985).Nucleotide sequence and expression in vitro of cDNA derived from mRNof int-1, a provirally activated mouse mammary oncogene. Mol. Cell Biol.5, 3337-3344.

Gan, L., Mao, C.-A., Wikramanayake, A., Angerer, L. M., Angerer, R. C.and Klein, W. H. (1995). An orthodenticle-related protein fromStrongylocentrotus purpuratus. Dev. Biol.167, 517-528.

Hemmati-Brivanlou, A., Kelly, O. G. and Melton, D. A. (1994). Follistatin,an antagonist of activin is expressed in the Spemann organizer and dispdirect neuralizing activity. Cell 77, 283-295.

Hermesz, E., Mackem, S. and Mahon, K. A. (1996). Rpx: a novel anterior-restricted homeobox gene progressively activated in the prechordal panterior neural plate and Rathke’s pouch of the mouse embryo. Development122, 41-52.

Hogan, B., Beddington, R., Constantini, F., and Lacy, E.(1994).Manipulating the Mouse Embryo: A Laboratory Manual, New York: ColdSpring Harbor.

Jollie, W. P. (1990). Development, morphology, and function of the yolk-saplacenta of laboratory rodents. Teratology41, 361-381.

Kaestner, K. H., Monaghan, A. P., Kern, H., Ang, S.-L., Weitz, S., Lichter,P. and Schutz, G. (1995). The mouse fkh-2gene. Implications for notochord,foregut, and midbrain regionalization. J. Biol. Chem.270, 30029-30035.

Kaufman, M. H. (1992). The Atlas of Mouse Development. New York:Academic Press.

Lemaire, P. and Kodjabachian, L. (1996). The verterbrate organizer:structure and molecules. Trends Genet.12, 525-530.

Li, Y., Allend, M. L., Finkelstein, R. and Weinberg, E. S. (1994). Expressionof zebrafish orthodenticle-related genes in the embryonic brain. MOD 48,229-244.

Matsuo, I., Kuratani, S., Kimura, C., Takeda, N. and Aizawa, S.(1995).

mal

ellyos.

-

an

n

cue

:

A

lays

late,

c

Mouse Otx2functions in the formation and patterning of rostral head. GenesDev. 9, 2646-2658.

McMahon, A. P., Joyner, A. L., Bradley, A. and McMahon, J. A.(1992).The midbrain-hindbrain phenotype of Wnt-1−/Wnt-1− mice results fromstepwise deletion of engrailedexpressing cells by 9.5 days postcoitum. Cell69, 581-595.

Mercier, P., Simeone, A., Coteli, F. and Boncinelli, E.(1995). Expressionpattern of the two otx genes suggests a role in specifying anterior bodystructures in zebrafish. Int. J. Dev. Biol.39, 559-573.

Millet, S., Bloch-Gallego, E., Simeone, A. and Alvarado-Mallart, R. M.(1996). The caudal limit of Otx2 gene expression as a marker of themidbrain/hindbrain boundary: a study using in situ hybridization andchick/quail homotopic grafts.Development122, 3785-3797.

Nagy, A., Rossant, J., Nagy, R., Abramov-Newely, W. and Roder.(1993).Viable cell culture-derived mice from early passage embryonic stem cellsProc. Natl. Acad. Sci. USA 90, 8424-8428.

Nornes, H. O., Dressler, G. R., Knapik, E. W., Deutsch, U. and Gruss, P.(1990). Spatially and temporally restricted expression of Pax2during murineneurogenesis. Development109, 797-809.

Oliver, G., Mailhos, A., Wehr, R., Copeland, N. G., Jenkins, N. A. andGruss, P. (1995). Six3, a murine homologue of thesine oculisgene,demarcates the most anterior border of the developing neural plate andexpressed during eye development. Development121, 4045-4055.

Pannese, M., Polo, C., Andreazzoli, M., Vignali, R., Kablar, B., Barsacchi,G. and Boncinelli, E. (1995). The Xenopus homologue of Otx2 is amaternal homeobox gene that demarcates and specifies anterior boregions. Development121, 707-720.

Robertson, E.J.(1987). Embryo-derived stem cell lines. In Teratocarcinomasand Embryonic Stem cells: a Practical Approach(ed E. J. Robertson),pp.71-112. Oxford: IRL Press.

Rowitch, D. H. and McMahon, A. P.(1995). Pax-2expression in the murineneural plate precedes and encompasses the expression domains of Wnt-1andEn-1. Mech. Dev.52, 3-8.

Sasai, Y., Lu, B., Steinbesser, H., Geissert, D., Gont, L. K., and DeRobertis, E. M. (1995). Xenopus chordin: a novel dorsalizing factoractivated by organizer-specific homeobox genes. Cell 79, 779-790.

Shawlot, W. and Behringer, R. R.(1995). Requirement for Lim1 in head-organizer function. Nature374, 425-430.

Simeone, A. Gulisano, M., Acampora, D., Stornaiuolo, A., Rambaldi, M.and Boncinelli, E. (1992). Two vertebrate homeobox genes related to theDrosophila empty spiraclesgene are expressed in the embryonic cerebracortex. EMBO J. 11, 2541-2550.

Simeone, A., Acampora, D., Mallamaci, A., Stornaiuolo. A., D’Apice. M.R., Nigro. V. and Boncinelli, E. (1993). A vertebrate gene related toorthodenticlecontains a homeodomain of the bicoid class and demarcateanterior neuroectoderm in the gastrulating mouse embryo. EMBO J. 12,2735-2747.

Smith, W. C. and Harland, R. M. (1992). Expression cloning of noggin, anew dorsalizing factor localized to the Spemann organizer in Xenopuembryos. Cell 70, 829-840.

Spyropoulus, D. D. and Capecchi, M. R.(1994). Targeted disruption of theeven-skippedgenes, evx1, causes early postimplantation lethality of themouse conceptus. Genes Dev.8, 1949-1961.

te Riele, H., Robanus Maandag, E., Clarke, A., Hooper, M. and Berns, A.(1990). Consecutive inactivation of both alleles of the pim-1proto-oncogeneby homologous recombination in embryonic stem cells. Nature348, 649-651.

Thomas, P. and Beddington, R.(1996). Anterior primitive endoderm may beresponsible for patterning the anterior neural plate in the mouse embryCurr. Biol. 6, 1487-1496.

Varlet, I., Collignon, J. and Robertson, E. J.(1997). nodal expression in theprimitive endoderm is required for specification of the anterior axis duringmouse gastrulation. Development124, 1033-1044.

Wilkinson, D. G., Bhatt, S., Cook, M., Boncinelli, E. and Krumlauf, R.(1989). Segmental expression of Hox-2homeobox genes in the developingmouse hindbrain. Nature341, 405-409.